jDUreport: Differential junction usage estimation
In chernolab/ASpli: Analysis of Alternative Splicing Using RNA-Seq

Description Usage Arguments Details Value Author(s) See Also Examples

This function estimates the differential usage of junctions combining different types of evidence

Differential junction usage is estimated using a combination of evidences

  jDUreport(asd, 
            minAvgCounts                       = 5, 
            contrast                           = NULL,
            filterWithContrasted               = TRUE,
            runUniformityTest                  = FALSE,
            mergedBams                         = NULL,
            maxPValForUniformityCheck          = 0.2,
            strongFilter                       = TRUE,
            maxConditionsForDispersionEstimate = 24,
            formula                            = NULL,
            coef                               = NULL,
            maxFDRForParticipation             = 0.05,
            useSubset                          = FALSE)

`asd`	An object of class `ASpliAS` with results of PSI and PIR using experimental junctions
`minAvgCounts`	Minimum average counts for filtering
`contrast`	Define the comparison between conditions to be tested. `contrast` should be a vector with length equal to the number of experimental conditions defined by `targets`. The values of this vector are the coefficients that will be used to weight each condition, the order of the values corresponds to the order given by `getConditions` function. When `contrast` is NULL, defaults to a vector containing -1, as the first value, 1 as the second an zero for all the remaining values, this corresponds to a pair comparison where the first condition is assumed to be a control and the second condition is the treatment, all other conditions are ignored. If NULL must provide a formula.
`filterWithContrasted`	A logical value specifying if bins, genes and junction will be filtered by read quantity and read density using data from those conditions that will be used in the comparison, i.e. those which coefficients in contrast argument are different from zero. The default value is TRUE, it is strongly recommended to do not change this value.
`runUniformityTest`	Run uniformity test on Intron Retention. Sometimes Mutually Exclusive Exons (MEX) events can be confused with Intron Retention events. This test compares the standard deviation of the inner intron region (11 bases from both ends) to the mean of both intron ends. Numbers closer to 0 mean the event is more probably an Intron Retention event than an MEX event. The test takes some time to run so it defaults to FALSE.
`mergedBams`	Path to merged bams for each testing condition. If no merged bams exist (for example, paired samples without replicates), use the same bams as targets.
`maxPValForUniformityCheck`	To speed up uniformity test only check junctions with pval < maxPValForUniformityCheck
`strongFilter`	If strongFilter is TRUE, then we remove all events with at least one junction that doesn't pass the filter.
`maxConditionsForDispersionEstimate`	In order to reduce resource usage, estimate dispersion for statistics tests with a reduced number of conditions.
`formula`	Either a formula or a contrast can be tested. If formula is used, complex tests can be run. `formula` should be a formula specifying which experimental conditions defined by `targets` to test. If coef is specified, then that coefficient will be tested. If not, it defaults to the last term in the formula.
`coef`	For formula only. The coefficient to be tested. If null the test defaults to the last term in the formula
`maxFDRForParticipation`	In order to calculate junctionPSI participation, only use significant junctions (ie junctions with FDR < maxFDRForParticipation).
`useSubset`	Experimental. It is strongly recomended to leave the default, FALSE.

Estimation is made at junction level using diffSpliceDGE function from edgeR package. Junctions belonging to the same AS event comprises the event "set". Each junction is tested against this "set" in a similar fashion that bins are tested against their gene in diffSpliceDGE. Localec are clusters made of junctions that share an end with at least another junction in the cluster.

An ASpliJDU object with results of differential usage at junctions level.

`localec`	size: number of junctions belonging to the cluster. cluster.LR: likelihood ratio of cluster differential usage. pvalue: pvalue of cluster differential usage. FDR: fdr of cluster differential usage. range: cluster location. participation: participation of the significant junction (FDR < maxFDRForParticipation) presenting maximal participation value inside the cluster dParticipation: delta participation of the significant junction (FDR < maxFDRForParticipation) presenting maximal participation value inside the cluster
`localej`	cluster: name of the cluster the junction belongs to log.mean: log of mean counts accross all conditions for this junction logFC: log fold change of junction accross conditions pvalue: pvalue of junction FDR: FDR of junction annotated: is junction annotated or new participation: the maximal participation value observed across contrasted condictions dParticipation: delta participation of the maximal participation value observed across contrasted condictions From dParticipation to the end, junction counts for all samples
`anchorc`	cluster.LR: likelihood ratio of cluster differential usage. pvalue: pvalue of cluster differential usage. FDR: fdr of cluster differential usage.
`anchorj`	log.mean: log of mean counts accross all conditions for this junction logFC: log fold change of junction accross conditions LR: likelihood ratio of junction differential usage. pvalue: pvalue of junction FDR: FDR of junction J1.pvalue: pvalue of J1 junction J2.pvalue: pvalue of J2 junction NonUniformity: if non uniformity test was performed, numbers closer to zero mean uniformity and closer to one mean non uniformity dPIR: junction delta PIR annotated: is junction annotated or new From annotated to the end, junction counts for all samples
`jir`	J3: J3 junction/s logFC: log fold change of junction accross conditions log.mean: log of mean counts accross all conditions for this junction pvalue: pvalue of junction FDR: FDR of junction LR: likelihood ratio of junction differential usage. NonUniformity: if non uniformity test was performed, numbers closer to zero mean uniformity and closer to one mean non uniformity dPIR: junction delta PIR multiplicity: do multiple junctions cross the region From multiplicity to the end, junction counts for all samples
`jes`	event: type of event J3: J3 junction/s logFC: log fold change of junction accross conditions log.mean: log of mean counts accross all conditions for this junction pvalue: pvalue of junction FDR: FDR of junction LR: likelihood ratio of junction differential usage. dPSI: junction delta PSI multiplicity: do multiple junctions cross the region From multiplicity to the end, junction counts for all samples
`jalt`	event: type of event J3: J3 junction/s logFC: log fold change of junction accross conditions log.mean: log of mean counts accross all conditions for this junction pvalue: pvalue of junction FDR: FDR of junction LR: likelihood ratio of junction differential usage. dPSI: junction delta PSI multiplicity: do multiple junctions cross the region From multiplicity to the end, junction counts for all samples
`contrast`	Conditions contrasted by ASpli

Estefania Mancini, Andres Rabinovich, Javier Iserte, Marcelo Yanovsky, Ariel Chernomoretz

Accesors: localec, localej, anchorc, anchorj, jir, jes, jalt, junctionsDU, Export: writeJDU, writeDU, edgeR, ASpliAS

  # Create a transcript DB from gff/gtf annotation file.
  # Warnings in this examples can be ignored. 
  library(GenomicFeatures)

  genomeTxDb <- makeTxDbFromGFF( system.file('extdata','genes.mini.gtf', 
                                 package="ASpli") )
  
  # Create an ASpliFeatures object from TxDb
  features <- binGenome( genomeTxDb )
  
  # Define bam files, sample names and experimental factors for targets.
  bamFileNames <- c( "A_C_0.bam", "A_C_1.bam", "A_C_2.bam", 
                     "A_D_0.bam", "A_D_1.bam", "A_D_2.bam" )

  targets <- data.frame( 
               row.names = paste0('Sample_',c(1:6)),
               bam = system.file( 'extdata', bamFileNames, package="ASpli" ),
               factor1 = c( 'C','C','C','D','D','D'),
               subject = c(0, 1, 2, 0, 1, 2))

  # Read counts from bam files
 gbcounts  <- gbCounts( features = features, 
                           targets = targets, 
                           minReadLength = 100, maxISize = 50000,
                           libType="SE", 
                           strandMode=0)
jcounts   <- jCounts(counts = gbcounts, 
                     features = features, 
                     minReadLength = 100,
                     libType="SE", 
                     strandMode=0)
                     

  
  # Test for factor1 controlling for paired subject
  jPaired <- jDUreport(jcounts, formula = formula(~subject+factor1))
  
  # Show junctions information
  jPaired
  localej(jPaired)
  localec(jPaired)
  anchorj(jPaired)
  anchorc(jPaired)
  jir(jPaired)
  jes(jPaired)
  jalt(jPaired)
  
  # Export results  
  writeJDU( jPaired, output.dir = paste0(tempdir(), "/jPaired") )